Integration of Meta-Multi-Omics Data Using Probabilistic Graphs and External Knowledge.

Area of Science:

• Microbiology
• Systems Biology
• Bioinformatics

Background:

• Multi-omics data offers a comprehensive view of biological systems, but analytical methods are underdeveloped.
• Kefir grains host a stable microbial community with known health benefits.
• Understanding microbial interactions is key to harnessing their potential.

Purpose of the Study:

• To develop and apply a novel algorithm for analyzing multi-omics data from bacterial communities.
• To construct a multifarious interaction network for *Lentilactobacillus kefiri* and *Lactobacillus kefiranofaciens* under different temperatures.
• To gain insights into microbial adaptation and community dynamics.

Main Methods:

• Developed a new algorithm using probabilistic graph representations and external knowledge for optimal structure learning.
• Collected transcriptomic, metabolomic, and proteomic data from controlled co-cultures of *L. kefiri* and *L. kefiranofaciens* at 30 °C and 37 °C.
• Constructed a multi-omics interaction network from 20 samples.

Main Results:

• The multi-omics network provided insights beyond single-omics analyses.
• Identified interactions between transcripts, proteins, and metabolites, suggesting active toxin/antitoxin systems.
• Observed complex interactions involving the shikimate pathway, explaining bacterial adaptation, co-aggregation, and differential activation of *L. kefiranofaciens* at 37 °C.

Conclusions:

• The developed algorithm effectively reconstructs complex microbial interaction networks from multi-omics data.
• The study elucidated molecular mechanisms underlying bacterial adaptation to temperature stress and community interactions.
• This approach enhances our understanding of microbial communities like those found in kefir.